AI Liability in Construction: Updating Your A201 for 2026

Artificial Intelligence is no longer a theoretical technology in construction. It’s embedded in tools for estimating, scheduling, safety monitoring, and design. Yet the contracts governing most projects, particularly the AIA A201™-2017 General Conditions, were written before this wave of AI adoption and remain silent on the issue. This silence is a significant, unaddressed risk.

When an AI tool produces a flawed design, an unworkable schedule, or a “hallucinated” specification, who is legally responsible? The answer, according to legal precedent and professional standards bodies like NCARB, is unequivocal: the licensed professional of record. The architect or engineer who stamps the drawing or the contractor who submits the schedule is accountable for the output, regardless of whether it was generated by a human or a machine. Software vendor license agreements almost universally disclaim liability for the outputs of their tools. You cannot outsource professional judgment or legal responsibility to an algorithm.

This guide provides a clause-by-clause framework for updating your AIA A201 contracts to address AI liability head-on. We will cover the primary risk scenarios, propose specific contractual language, and review tools where these compliance issues are most acute. ZEKAI reviews all tools independently; our goal is to equip professionals in the construction and engineering fields to adopt technology responsibly.

Why Standard AIA Contracts Fail to Cover AI Risk

The AIA A201™-2017, the industry’s most-used contract document, is a robust framework for allocating risk—but it never mentions “Artificial Intelligence.” This omission creates critical gaps. The core problem lies with the “standard of care,” the legal duty for professionals to perform their work with the skill and judgment a reasonably prudent peer would exercise. AI complicates this in two ways:

1. Opacity: Many AI tools are “black boxes,” making it impossible to audit *how* they reached a conclusion. If you can’t explain the logic, can you claim you exercised professional judgment?
2. Novel Errors: AI can produce confident, plausible-sounding errors—like citing non-existent building codes or specifying materials that are physically incompatible—that a human expert might not make.

While the AIA has published digital practice documents like the E203™-2013 (now superseded by 2022 documents), these were primarily designed for BIM and digital data exchange, not the predictive and generative capabilities of modern AI. They provide a starting point for data protocols but do not adequately address liability for AI-generated design, analysis, or site monitoring. The responsibility for adapting contracts falls to the parties on each project.

The Three Core AI Liability Scenarios

Before drafting new clauses, it’s essential to understand the specific risks. We see three primary categories of AI-related disputes emerging in construction.

1. Design and Preconstruction Errors

Generative design tools can produce thousands of design options, and AI-powered takeoff software can estimate quantities in minutes. But these tools can also “hallucinate” unbuildable geometry, misinterpret drawings, or base estimates on flawed data. If an architect incorporates an AI-generated, code-non-compliant design element into stamped drawings, the architect is liable. If a contractor submits a bid based on an AI’s erroneous quantity takeoff, the contractor owns the financial consequences of that error.

2. Scheduling and Management Failures

AI scheduling platforms can analyze millions of sequences to optimize a project’s critical path. However, their effectiveness depends entirely on the quality of the input data. If a schedule is generated from incomplete historical data or fails to account for a site-specific constraint, the resulting “optimized” plan can lead to significant delays and disputes. The contractor who submits that schedule as their plan of record is ultimately responsible for its viability.

3. Jobsite Monitoring and Data Privacy Violations

AI-powered camera systems can track progress, spot safety hazards, and verify worker attendance. This creates two distinct liability paths. First, if the system fails to flag a genuine safety hazard that leads to an incident, the contractor may face questions about negligent reliance on the technology. Second, and more immediately, many of these systems use biometric identifiers like facial scans for timekeeping. This directly implicates strict privacy laws like Illinois’ Biometric Information Privacy Act (BIPA), which mandates specific written consent and data-handling policies. A violation can lead to statutory damages per employee, per violation, creating massive financial exposure.

How to Update Your AIA A201: A Clause-by-Clause Guide

Amending the A201 should be done deliberately and with legal counsel. The goal is not to disclaim all AI risk, but to define its use and allocate responsibility clearly. Parties should consider creating a project-specific “AI Use Protocol” exhibit, similar to a BIM Execution Plan, that is incorporated by reference.

Here are four key areas of the A201 to address as of September 2026.

Swipe the table sideways →

You are a construction lawyer. Draft a preliminary contract clause for an addendum to an AIA A201-2017 agreement. The clause should define "Artificial Intelligence (AI) Tool," establish that each party is responsible for its own use of AI tools, and require that any party using an AI tool to generate an output (e.g., a design, schedule, or report) must have that output independently verified by a qualified human professional before it is submitted or relied upon by other parties. The clause must affirm that the professional standard of care applies to all work, regardless of whether AI was used as a tool.

Featured AI Tools with Compliance Considerations

The contractual risks are not abstract. They apply directly to tools being deployed on jobsites today. We review hundreds of tools at ZEKAI, and for this topic, two platforms highlight the intersection of innovation and liability.

The Path Forward: Adopt, but Verify and Allocate

AI tools offer immense potential to make construction more efficient, accurate, and safe. However, adopting the technology without updating the contractual frameworks that govern our projects is a recipe for disputes. The legal and professional consensus is clear: a human expert must stand behind the work.

The core principle is to treat AI output as raw material requiring professional judgment, not as finished work product. Before deploying any AI tool on a project, all parties—owners, architects, and contractors—must have a frank discussion about risk. That conversation should be documented in the A201 and any corresponding agreements. By defining the rules of engagement before a problem arises, teams can put AI to work while protecting themselves from the significant liabilities it can create.

To learn more about the tools transforming the industry, visit our hub for AI in construction and engineering.